Year: 2025
Type: article
Abstract: The high volatility and complexity of the stock market make price prediction a challenging time series problem. Traditional methods such as ARIMA perform well in capturing linear relationships but struggle to effectively address the nonlinear and non-stationary characteristics of stock prices. In recent years, Long Short-Term Memory (LSTM) networks have garnered widespread attention due to their ability to handle sequential data and long-term dependencies. However, a standalone LSTM model is still insufficient when dealing with the high nonlinearity and noise inherent in stock data. To address this, this paper proposes a hybrid model combining Ensemble Empirical Mode Decomposition (EEMD) and LSTM to improve prediction accuracy and stability. By decomposing stock price time series using EEMD, intrinsic mode functions (IMFs) and residual components of different frequencies are extracted. These components are then individually predicted using LSTM models, and the results are weighted and integrated to generate the final prediction. Experimental results demonstrate that the proposed hybrid model significantly outperforms traditional methods in both prediction accuracy and robustness, showcasing its important application value in stock market forecasting.
DOI: 10.54097/hqx9sy23
Authors: Jianliang Zhang, Qirong Zhang
Institutions: Chongqing Jiaotong University, Wuhan University of Technology
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W4409457583